[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250829171855.64f2cbfc@gandalf.local.home>
Date: Fri, 29 Aug 2025 17:18:55 -0400
From: Steven Rostedt <rostedt@...dmis.org>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Arnaldo Carvalho de Melo <arnaldo.melo@...il.com>, Steven Rostedt
<rostedt@...nel.org>, linux-kernel@...r.kernel.org,
linux-trace-kernel@...r.kernel.org, bpf@...r.kernel.org, x86@...nel.org,
Masami Hiramatsu <mhiramat@...nel.org>, Mathieu Desnoyers
<mathieu.desnoyers@...icios.com>, Josh Poimboeuf <jpoimboe@...nel.org>,
Peter Zijlstra <peterz@...radead.org>, Ingo Molnar <mingo@...nel.org>, Jiri
Olsa <jolsa@...nel.org>, Arnaldo Carvalho de Melo <acme@...nel.org>,
Namhyung Kim <namhyung@...nel.org>, Thomas Gleixner <tglx@...utronix.de>,
Andrii Nakryiko <andrii@...nel.org>, Indu Bhagat <indu.bhagat@...cle.com>,
"Jose E. Marchesi" <jemarch@....org>, Beau Belgrave
<beaub@...ux.microsoft.com>, Jens Remus <jremus@...ux.ibm.com>, Andrew
Morton <akpm@...ux-foundation.org>, Florian Weimer <fweimer@...hat.com>,
Sam James <sam@...too.org>, Kees Cook <kees@...nel.org>, "Carlos O'Donell"
<codonell@...hat.com>
Subject: Re: [PATCH v6 5/6] tracing: Show inode and device major:minor in
deferred user space stacktrace
On Fri, 29 Aug 2025 13:54:08 -0700
Linus Torvalds <torvalds@...ux-foundation.org> wrote:
> On Fri, 29 Aug 2025 at 11:11, Steven Rostedt <rostedt@...dmis.org> wrote:
> >
> > The idea is this (pseudo code):
> >
> > user_stack_trace() {
> > foreach vma in each stack frame:
> > key = hash(vma->vm_file);
> > if (!lookup(key)) {
> > trace_file_map(key, generate_path(vma), generate_buildid(vma));
> > add_into_hash(key);
> > }
> > }
>
> I see *zero* advantage to this. It's only doing stupid things that
> cost extra, and only because you don't want to do the smart thing that
> I've explained extensively that has *NONE* of these overheads.
>
> Just do the parsing at parse time. End of story.
What does "parsing at parse time" mean?
>
> Or don't do this at all. Justy forget the whole thing entirely. Throw
> the patch that started this all away, and just DON'T DO THIS.
Maybe we are talking past each other.
When I get a user space stack trace, I get the virtual addresses of each of
the user space functions. This is saved into an user stack trace event in
the ring buffer that usually gets mapped right to a file for post
processing.
I still do the:
user_stack_trace() {
foreach addr each stack frame
vma = vma_lookup(mm, addr);
callchain[i++] = (addr - vma->vm_start) + (vma->vm_pgoff << PAGE_SHIFT);
Are you saying that this shouldn't be done either? And to just record the
the virtual address in the chain and the vma->vm_start and
vma->vm_pgoff in another event? Where the post processing could do the
math? This other event could also record the path and build id.
The question is, when do I record this vma event? How do I know it's new?
I can't rely too much on other events (like mmap) and such as those events
may have occurred before the tracing started. I have to have some way to
know if the vma has been saved previously, which was why I had the hash
lookup, and only add vma's on new instances.
My main question is, when do I record the vma data event?
-- Steve
Powered by blists - more mailing lists